Alzheimer's disease (AD) is a prevalent neurodegenerative disorder among elderly population, with over 5 millions of AD patients in the US and the social and economic cost exceeding 200 billion dollars in 2012. However, currently there is no effective treatment available for AD. Reactive astrocytes are an important pathological marker widely associated with neural injury and neurodegenerative disorders including AD. Astrogliosis is a major obstacle to prevent neuronal axon regeneration and axon re- entry into the injury or diseased areas. We have developed an innovative approach to trans- differentiate reactive astrocytes into functional neurons in AD mouse brain in vivo. This was achieved through the use of retrovirus to overexpress proneural transcription factor NeuroD1 in a transgenic AD mouse model (5xFAD). This proposal is to further evaluate this novel approach as a potential therapy to treat AD. The specific aims are: 1) To investigate local network remodeling in AD brains after converting reactive astrocytes into functional neurons. 2) To generate novel transgenic mouse models for efficient reactive astrocyte-neuron conversion. 3) To examine functional and behavioral improvement after reactive astrocyte-neuron conversion in AD brains. Our in vivo trans-differentiation approach to convert reactive astrocytes into functional neurons not only reduces the number of reactive astrocytes, but also generates new neurons in diseased areas for brain repair. This internal trans-differentiation strategy will avoid potential aversive effects associated with cell transplantation method. The trans-differentiation technology described in this proposal may be potentially developed into novel therapeutic treatment to reverse astrogliosis induced by brain injury or neurodegenerative disorders.